I never actually figured out exactly what was wrong in entry 2162, but I managed to circumvent by changing the time sequence of events in the up script, moving the big gain increases in the common mode servo to the end of the script.  So the IFO can be locked again.